7742 • The Journal of Neuroscience, August 27, 2003 • 23(21):7742–7749
Cellular/Molecular
Differentiation of Marrow Stromal Cells into Photoreceptors
in the Rat Eye
Anthony Kicic,1 Wei-Yong Shen,2 Ann S. Wilson,3 Ian J. Constable,2 Terry Robertson,4 and P. Elizabeth Rakoczy2
Stem Cell Unit, Department of Molecular Ophthalmology, Lions Eye Institute, affiliated with the Center of Ophthalmology and Visual Science, University of
Western Australia, Nedlands, 6009, Western Australia, Australia, 2Center of Ophthalmology and Visual Science, University of Western Australia, Nedlands,
6009, Western Australia, Australia, 3Department of Molecular Ophthalmology, Lions Eye Institute, affiliated with the Center of Ophthalmology and Visual
Science, University of Western Australia, Nedlands, 6009, Western Australia, Australia, and 4Department of Pathology, The University of Western
Australia, Nedlands, 6009, Western Australia, Australia
1
Retinal degenerations and dystrophies are the major causes of genetically inherited blindness that are characterized by the apoptotic
death of the photoreceptor cell layer of the retina. To date, no treatment exists for these diseases and only recently have they been
considered as candidates for gene and stem cell therapies. Here we report the ability of adult CD90 ⫹ marrow stromal cells (MSCs) to be
induced by activin A, taurine, and EGF into cells (20 –32%) expressing photoreceptor-specific markers rhodopsin, opsin, and recoverin
in vitro. CD90 ⫹ cells were either transduced with recombinant adeno-associated virus expressing green fluorescent protein (GFP) or
bromodeoxyuridine (BrdU) labeled and then injected into the subretinal space of adult Royal College of Surgeons rats. Fundus photography and angiography showed no adverse effects of CD90 ⫹ MSC transplantation. GFP-expressing cells or BrdU-positive cells covered
⬃30% of the entire retinal area. By 2 weeks after injection, CD90 ⫹ MSCs integrated into the host retina, forming structures similar to the
photoreceptor layer and expressed a photoreceptor-specific marker. No teratoma formation was observed in the recipient retina. The
subretinally delivered CD90 ⫹ MSCs did not stain for proliferating cell nuclear antigen, indicating that they primarily undergo differentiation rather than proliferation. In addition, we established that transplanted cells can attract synaptic vesicles and hence are potentially
capable of signal transduction. This study demonstrates for the first time the partial differentiation of adult CD90 ⫹ MSCs into photoreceptors in vitro and in vivo. Our results establish a proof of concept for CD90 ⫹ MSC differentiation with autologous transplantation,
which may provide a promising therapeutic strategy for the treatment of some forms of genetically inherited retinal degenerations.
Key words: marrow stromal cells; photoreceptors; retinal degeneration; stem cells; cell-based therapy; plasticity
Introduction
Retinal degenerations and dystrophies are the major causes of
genetically inherited blindness in the developed world. These
conditions are typically characterized by the apoptotic death of
one of a subset of cells in the retina, namely the photoreceptors
(Gavrieli et al., 1992; Lolley et al., 1994). Because visually evoked
responses have been recorded in animals suffering degeneration
after photoreceptor transplantation (Huang et al., 1998; Woch et
al., 2001), suggesting that despite the partial or complete loss of
photoreceptors, function may be maintained in the inner nuclear
layer and the axons connecting the retina to the brain. Hence,
photoreceptor replacement in the form of a cell-based therapeutic approach might possibly aid in the restoration of some degree
of vision. Embryonic stem cell therapy, although promising, remains highly controversial and must overcome the potential
problem of rejection. In this respect, if an accessible pluripotent
Received Aug. 26, 2002; revised June 23, 2003; accepted June 24, 2003.
We thank Prof. Miranda Grounds for inspirational discussions. We sincerely thank Dr. Matthew Wikstrom, Christine Hall, Tammy Zaknich, Dr. Meliha Brankov, and Ben Rae for their technical assistance. We also thank Dr. Elizabeth
Kicic-Starcevich for her valuable constructive comments.
Correspondence should be addressed to Dr. Anthony Kicic, Stem Cell Unit, Department of Molecular Ophthalmology, Lions Eye Institute, affiliated with the Center of Ophthalmology and Visual Science, University of Western
Australia, 2 Verdun Street, Nedlands, 6009, Western Australia, Australia. E-mail: [email protected].
Copyright © 2003 Society for Neuroscience 0270-6474/03/237742-08$15.00/0
stem cell source can be identified, autologous stem cell therapies
offer a great advantage.
A limited number of retinal stem cells have been successfully
extracted from the pigment ciliary margin (Tropepe et al., 2000)
and differentiated into neurons, such as photoreceptors, in vitro.
Extraction of these cells involves complicated microsurgical procedures, and the limited availability of pluripotent retinal stem
cells means that this approach is not an option for the treatment
of genetically inherited ocular diseases. In contrast, the bone
marrow is an ideal source of pluripotent stem cells, because it is
the primary site of hematopoietic stem cell renewal and differentiation. It is comprised of at least two types of stem cells: hematopoietic stem cells and stem cells for nonhematopoietic tissues,
which commonly referred to as marrow stromal cells (MSCs).
These highly undifferentiated, self-renewing elements were originally thought to be severely limited in their capacity for differentiation, but recent reports of their differentiation into muscle,
glial, hepatic, renal, and neural lineages has challenged this notion (Eglitis and Mezey, 1997; Ferrari et al., 1998; Petersen et al.,
1999; Kopen et al., 1999). Because of its autologous characteristic,
relative ease of isolation, and its less controversial nature, this
pool of pluripotent stem cells remains a forerunner as the cells of
choice in the treatment of diseases using cell-based therapy.
A number of investigations have shown the multilineage po-
Kicic et al. • Marrow Stromal Cell Differentiation into Photoreceptors
tential of MSCs into neural lineages (Azizi et al., 1998; Brazelton
et al., 2000; Mezey et al., 2000), which prompted us to attempt
their induction into photoreceptors in vitro and in vivo. For the
first time, we show the ability of adult CD90 ⫹ MSCs to be induced into cells expressing photoreceptor-specific markers in
vitro using activin A, taurine, and epidermal growth factor (EGF).
In addition, and more significantly, we demonstrate the successful transplantation of adult CD90 ⫹ MSCs into the adult rat eye
with minimal side effects, the in vivo survival of these cells, their
partial integration and differentiation into the host retina, and
potential functional activity.
Materials and Methods
Reagents. ␣ minimal essential medium (␣MEM), fetal bovine serum
(FBS), normal rabbit serum, penicillin G, streptomycin sulfate, trypsinEDTA, L-glutamine, transforming growth factor ␤1 (TGF␤1), murine
natural EGF, and TRIzol reagent were purchased from Invitrogen (Melbourne, Australia). Insulin, 5,8,11,14-eicosatetraynoic acid, dexamethasone, bovine serum albumin (BSA), activin A, and taurine were obtained
from Sigma (St Louis, MO). PharMingen (San Diego, CA) supplied all
primary antibodies and the streptavidin-allophycocyanin conjugate
(Sav-APC), with the exception of anti-opsin and anti-collagen type II and
all secondary antibodies mentioned, which were obtained from Sigma.
The recoverin antibody was purchased from ProteinTech Group (Chicago, IL). Triton X-100 was supplied by BDH Chemicals (Victoria, Australia). All tissue culture plasticware was obtained from Becton Dickinson
(Franklin Lakes, NJ).
Isolation and culture of MSCs. The primary culture of rat MSCs was
performed using standard techniques described previously (Phinney et
al., 1999; Woodbury et al., 2000), in accordance with the policy of The
Society of Neuroscience for animal use in experimentation.
Stromal cell characterization and purification. The CD90 ⫹ MSCs were
analyzed using fluorescent-activated cell sorting (FACS) (FACSCalibur;
Becton Dickinson). In brief, harvested cells were fixed with 70% methanol, washed in PBS, and incubated for 1 hr on ice with fluorescentconjugated monoclonal antibodies (1:100 dilution) directed against the
following cell surface markers: CD11b, CD45, and CD90. After washing,
secondary antibody (SAv-APC; 1:100 dilution) was added, and the cells
were incubated on ice for 1 hr. Cells were washed an additional two times
before analysis. For purification, the cells were sorted using a FACSVantage cell sorter (Becton Dickinson) after the fluorescence-labeling procedure mentioned above, with the exception that the cells were not fixed
before analysis. MSCs solely expressing CD90 were then aseptically collected, recultured, and used in all subsequent experiments.
Multipotency of CD90⫹ MSCs. Adipogenic differentiation was performed using a method described previously (Kopen et al., 1999), and the
cells were stained with Oil Red O and Sudan Black B. Chondrogenic
differentiation was performed using a modified method of Muraglia et al.
(2000), in which CD90 ⫹ cells were cultured in ␣MEM and TGF␤1 (10
ng/ml) for 8 –10 d. Cells were then fixed and stained for type II collagen
expression.
Photoreceptor induction in vitro. After 24 hr in culture, basal medium
for CD90 ⫹ cells was replaced with ␣MEM supplemented with FBS (1%
v/v), penicillin (100 U/ml), streptomycin (100 ␮g/ml), and one of the
following inducing agents: activin A (100 ng/ml), taurine (50 ␮M), or
EGF (100 ng/ml). Cells were cultured for an additional 8 –10 d, fixed, and
then immunocytochemically analyzed.
Immunocytochemistry. CD90 ⫹ MSCs were fixed, washed, and then
blocked in 5% (w/v) BSA, 10% FBS (v/v), and 0.1% (v/v) Triton X-100 in
1⫻ PBS for 1 hr at room temperature. Cells were incubated with various
primary antibodies for 24 hr at 4°C followed by secondary antibodies
(FITC–tetramethylrhodamine isothiocyanate-conjugated or Streptavidin biotin) for a similar period. The antibody complex was visualized
using a fluorescent microscope (Olympus Optical, Tokyo, Japan). Primary antibodies included the following: microtubule-associated protein
2 (MAP2; 1:500), glial fibrillary acid protein (GFAP; 1:500), rhodopsin
(RHOS; 1:100), protein kinase C (PKC; 1:100), cellular retinoic acidbinding protein-1 (CRABP1; 1:100), and nestin (1:1000).
J. Neurosci., August 27, 2003 • 23(21):7742–7749 • 7743
Western blot analysis. CD90 ⫹ MSCs incubated with each inducing
agent over a 10 d period were collected and lysed, and ⬃40 ␮g of cellular
protein as well as rat retinal tissue (positive control) was electrophoresed
on a 12% (w/v) SDS-polyacrylamide gel. Separated proteins were transferred electropheretically onto a nitrocellulose membrane (100 mA; 1 hr;
4°C) and immunoblotted using photoreceptor-specific antibodies. In
brief, membranes were blocked for 1 hr at room temperature with 5%
(w/v) skim milk powder in PBS and 0.05% (v/v) Tween 20. After washing, membranes were incubated at room temperature with primary antibodies [recoverin, 1:100 diluted in 5% (w/v) skim milk powder in PBS;
opsin, 1:500 diluted in PBS] for 1 hr and then washed three times at 10
min intervals. The membranes were then incubated with 1:20,000 dilution of the secondary anti-rabbit IgG and anti-mouse IgG horseradish
peroxidase-linked antibodies for recoverin and opsin, respectively, for 1
hr at room temperature and then washed for another 30 min. Opsin and
recoverin expression was then visualized using the ECL and ECL Plus
Western blotting detection systems, respectively (Amersham Biosciences, Buckinghamshire, UK).
Semiquantitative reverse transcriptase PCR of the genes for differentiation markers of multilineages. Total RNA was extracted from the cells
using TRIzol reagent and was enriched for mRNA using the RNeasy mini
columns (Qiagen, Hilden, Germany) following the instructions of the
manufacturer. Primers were designed for the following genes:
glyceraldehydes-3-phosphate dehydrogenase (GAPD) (GenBank accession number AF106860) 5⬘(301–321), 3⬘(561–542); PCNA ( proliferating
cell nuclear antigen) (GenBank accession number NM 022381) 5⬘(163–
182), 3⬘(1116 –1094); recoverin (GenBank accession number D12573)
5⬘(429 – 449), 3⬘(1440 –1420); peripherin (GenBank accession number
AF031878) 5⬘(671– 691), 3⬘(1540 –1520); opsin (GenBank accession
number U22180) 5⬘(73–93), 3⬘(504 – 485); lipoprotein lipase (GenBank
accession number L03294) 5⬘(2811–2831), 3⬘(3468 –3448); and collagen
type II (GenBank accession number L48440) 5⬘(3509 –3527), 3⬘(3982–
3962). The cDNA was synthesized and amplified using the OneStep reverse transcriptase (RT)-PCR kit (Qiagen) under the following conditions: reverse transcription at 50°C for 30 min, initial PCR activation at
95°C for 15 min followed by 35 cycles of denaturation at 94°C for 1 min,
annealing at 56 – 62°C for 1 min, and extension at 72°C for 1 min. A final
extension step at 72°C for 10 min was included before samples were being
cooled to 4°C. The RT-PCR samples were then separated by 1.6% agarose
gel electrophoresis and photographed.
In vivo transplantation of CD90⫹ MSCs. A total of 28 Royal College of
Surgeons (RCS) rats were subretinally injected with stem cells using a
modified method of Shen and Rakoczy (2001). Between 50 and 100,000
CD90 ⫹ MSCs prelabeled with either bromodeoxyuridine (BrdU; 5 ␮M)
or recombinant adeno-associated virus expressing green fluorescent protein (rAAV.GFP) in 2 ␮l of PBS were injected into the left eye while the
other eye (control) was injected with 2 ␮l of PBS. Eyes were monitored up
to 12 weeks after transplantation by fundus color photography and angiography. Enucleated eyes were then frozen in optimal cutting temperature compound, sectioned, and then stained for BrdU (1:50), rhodopsin
(1:100), synaptophysin (1:200), and PCNA (1:250) using the methods
described above. Control sections were counterstained to observe nuclei.
For electron microscopic examination, animals were killed under halothane anesthetic and eyes were carefully removed and placed into a fixative solution containing 4% paraformaldehyde and 0.5% glutaraldehyde
in 0.05 M cacodylate buffer, pH 7.4. The eyes were then trimmed and
reimmersed into a fresh solution of the same fixative for an additional 24
hr. After postfixing in 1% osmium tetroxide, the tissues were then processed for transmission electron microscopy (TEM) by conventional
methods and embedded in Araldite. Semithin sections (1 ␮m thick) were
stained with 0.5% toluidine blue in 5% borax and examined with a Zeiss
(Thornwood, NY) light microscope. Immunohistochemical analysis using BrdU was used to identify transplanted cells. Staining revealed that a
majority of transplanted cells was found at the injection site or at the
proximal end of the outer nuclear layer (ON) (data not shown). These
sections, in which transplanted cells predominated, were then carefully
trimmed under a dissecting microscope, and thin sections (70nm thick)
were prepared using an LKB–Wallac (Gaithersburg, MD) Nova ultramicrotome. The sections were stained with Reynold’s lead citrate and ex-
Kicic et al. • Marrow Stromal Cell Differentiation into Photoreceptors
7744 • J. Neurosci., August 27, 2003 • 23(21):7742–7749
Table 1. The effect of passage on MSC differentiation
Passage number
CD45 (%)
CD11b (%)
CD90 (%)
1
4
8
12
15
20
23
⫺
⫺
⫺
⫺
⫺
⫺
⫺
4.15 ⫾ 1.7
2.11 ⫾ 0.83
1.98 ⫾ 0.57
1.25 ⫾ 0.24
0.61 ⫾ 0.02
0.83 ⫾ 0.2
0.1 ⫾ 0.01
84 ⫾ 3.8
80.3 ⫾ 4.4
68.92 ⫾ 6.1
66.45 ⫾ 7.8
41.75 ⫾ 6.2
42.22 ⫾ 5.9
17.5 ⫾ 5.5
The percentage of MSCs expressing CD90 remains constant up to passage four but significantly decreases (p ⬍
0.001) by the eighth passage.
after eight passages, indicating that an accessible source of undifferentiated cells exists in early passage MSCs.
Figure 1. A–C, Forward and side scatter plots of CD45 ( A), CD11b ( B), and CD90 ( C) cells
showing the characterization of undifferentiated rat MSCs. Population profiles of first-passage
MSCs were found to primarily express CD90 (84 ⫾ 3.8%).
amined in a Philips (Einhoven, The Netherlands) 410LS TEM at an accelerating voltage of 80 kV.
Transduction of CD90⫹ MSCs. CD90 ⫹ MSCs were transduced with an
rAAV.GFP in ␣MEM using the method of Lai et al. (2001).
Results
Stromal cell characterization
With the aim of producing a more unified cell population, primary cultures of MSCs obtained from normal adult RCSrdy ⫹-p ⫹ rats were passaged once before analysis by fluorescent
flow cytometry. We analyzed three individually prepared samples
of MSCs and found all cells to be negative for CD45, the cell
surface marker for the leukocyte cell lineage (Fig. 1 A, Table 1). In
addition, a small proportion of the population (2.4 –5.8%) was
found to express CD11b, the standard cell surface marker for the
erythrocyte lineage (Fig. 1 B, Table 1). In contrast, ⬃84% of firstpassage MSCs expressed CD90 (Fig. 1C, Table 1), the undifferentiated cell marker. This population profile changed markedly
with continual passage (Table 1). MSCs of early passage primarily
expressed CD90 (⬎80%), which decreased with subsequent passages to ⬍20% (Table 1). However, a statistically significant decrease ( p ⬍ 0.001) in the number of CD90 ⫹ cells only occurred
In vitro photoreceptor differentiation
The plasticity of CD90 ⫹ MSCs was confirmed by initiating their
differentiation into adipogenic and chondrogenic cell lineages.
This was confirmed by RT-PCR, demonstrating the expressions
of lipopolylipase and type II collagen, respectively (data not
shown).
The potential of CD90 ⫹ MSCs to differentiate into any of the
retinal cell lineages was assessed by exposing them to the following specific inducing agents: taurine, activin A, and EGF. After
induction with taurine, activin A, and EGF, ⬃30% of the cell
population expressed RHOS, a photoreceptor-specific cell
marker (Figs. 2 B--D, 3A). Similarly, 30% of the CD90 ⫹ MSCs
stained positively for CRABP1, an amacrine cell marker (Figs.
2G--I, 3A). All three agents also induced a small number of
CD90 ⫹ MSCs (⬃6 –10%) to express nestin, a typical neural stem
cell marker (Figs. 2 L--N, 3A). Untreated control CD90 ⫹ MSCs
did not stain against any of the markers tested (Fig. 2 E, J,O). In
addition, none of the treated MSCs showed positive staining with
the monoclonal antibody to PKC, which specifically identifies
bipolar cells in the retina, astrocytic marker GFAP, or MAP2, a
neural cell marker (data not shown).
Photoreceptor-specific differentiation was also confirmed using Western blot analysis and RT-PCR. Incubation of CD90 ⫹
MSCs with activin A, taurine, or EGF (Fig. 3B, lanes 3–5, C,D,
lanes 2– 4, respectively) all resulted in the expression of the
photoreceptor-specific encoders opsin (Fig. 3 B, C) and recoverin
(Fig. 3 B, D). Undifferentiated control CD90 ⫹ MSCs incubated in
a basal medium did not express any of these markers (Fig. 3B,
lane 2, C,D, lane 5). CD90 ⫹ MSCs incubated in the same inducing agents (Fig. 3E, lanes 6 – 8) as well as undifferentiated controls
(Fig. 3E, lane 9) were all found to express the proliferation marker
PCNA. Equal loading of samples was confirmed by the expression of GAPD (Fig. 3E, lanes 2–5).
In vivo photoreceptor differentiation
Having established that CD90 ⫹ MSCs could be driven to express
photoreceptor lineage-specific markers in vitro, we examined the
influence of the ocular microenvironment on the fate of these
cells. Subretinal injection of both PBS and a single-cell suspension of uninduced CD90 ⫹ MSCs into 4 –5-week-old recipient
animals resulted in the formation of a subretinal bleb covering
approximately one-third of the entire retina (Fig. 4 A, B). The
injected subretinal fluid was almost entirely absorbed by day 4
after injection (data not shown). Color fundus photography between 1 and 5 weeks after injection revealed no teratoma formation in the recipient retina (Fig. 4C,E). Immunohistochemical
analysis also confirmed the nonproliferating nature of transplanted CD90 ⫹ MSCs by the lack of staining for the cell prolif-
Kicic et al. • Marrow Stromal Cell Differentiation into Photoreceptors
J. Neurosci., August 27, 2003 • 23(21):7742–7749 • 7745
the retinal bleb created initially (Fig.
4C,E). No neovascularization, leakage, or
other vascular changes were observed after
CD90 ⫹ MSC transplantation (Fig. 4G,H ).
Immunohistochemistry of retinas containing transplanted BrdU-labeled (Fig.
5A) CD90 ⫹ MSCs confirmed that the
transplanted cells were distributed around
the injection site and remained present up
to 12 weeks (duration of observation) (Fig.
5B--E). Subretinal injection of MSCs resulted only in a slight disruption to the cellular architecture in the host retinas (Fig.
5 B, C,E), which was localized primarily at
the site of injection. Within the original
bleb, little or no disruption to host retinas
was observed at sites distant to the injecFigure 2. A, F, K, Phase contrast micrographs of CD90 ⫹ MSCs induced by activin A. Fluorescent micrographs of immunocyto- tion site (Fig. 5D). There was also a sub⫹
chemical analysis of CD90 MSCs induced by activin A ( B, G, L), taurine (C, H, M ), and EGF ( D, I, N ), using photoreceptor (RHOS)
stantial migration of BrdU-labeled cells
( B–E), amacrine (CRABP1) ( G–J), and neural (nestin) ( L–O) specific antibodies, are shown. Uninduced CD90 ⫹ MSCs cells did not
across the retina within a short period of
stain for any markers tested ( E, J, O). Magnification, 400⫻.
time. Within 2 weeks after injection, a
number of CD90 ⫹ MSCs moved across
the photoreceptor outer and inner segment layers and incorporated into the ON among the nuclei of photoreceptors in the area
defined by the bleb (Fig. 5G). The presence of BrdU-labeled cells
in the ON was also confirmed at 12 weeks after injection, although signal intensity was weaker (Fig. 5I ). There were no
BrdU-labeled cells found in areas distant to the injection site (Fig.
5 F, H ). These results demonstrated that CD90 ⫹ MSCs not only
integrated into the retina but also remained there for an extended
period of time, providing an opportunity for these cells to differentiate into neural retinal cells.
At all time points, the majority of the integrated BrdU-labeled
cells was found in the ON (Fig. 5J ). Mature photoreceptors have
a unique structure comprising of stacks of photoreceptor nuclei
forming the outer nuclear layer and an elongated cytoplasm
forming the photoreceptor inner and outer segments (Besharse
and Pfenninger, 1980). Opsin, the protein forming the visual
pigment, is normally present in the photoreceptor inner and
outer segments but is not detectable in the outer nuclear layer
(Fig. 5K ). Eyes injected with CD90 ⫹ MSCs show the typical outer
segment staining of opsin in undisrupted retinas distant to the
injection site (Fig. 5L). However, the vast majority of cells located
at the injection site demonstrated typical RHOS antibody staining in their cytoplasm (Fig. 5 M, N ), which paralleled observations made in vehicle-injected control animals (data not shown).
Additional investigation demonstrated that a number of the cells
that were immunoreactive for the RHOS antibody within their
cytoplasm were also positive for BrdU (Fig. 5O, bottom arrows)
and could be distinguished easily from endogenous photoreceptors, the opsin expression of which had been translocated to their
⫹
Figure 3. A, Quantification of CD90 cell differentiation after induction with activin A taucytoplasm caused by the retinal detachment resulting from the
rine or EGF using RHOS (photoreceptor), CRABP1 (amacrine), and nestin (neural)-specific mark⫹
injection procedure (Fig. 5O, top arrow). It was interesting to
ers, respectively. B, Western blot analysis of induced CD90 MSCs. Note that rat retinal tissue
note that whereas BrdU demonstrated nuclear localization,
(lane 1), undifferentiated CD90 ⫹ MSCs (lane 2), activin A (lane 3), taurine (lane 4), and EGF
RHOS staining was cytoplasmic surrounding the green nuclei
(lane 5) are shown. C, D, RT-PCR of opsin ( C) and recoverin ( D). Note that, in C–E, activin A (lanes
2, 6), taurine (lanes 3, 7), EGF (lanes 4, 8), and undifferentiated CD90 ⫹ cells (lanes 5, 9) are
(Fig. 5O, inset). However, even at 12 weeks after injection, the
shown. C, D, Undifferentiated CD90 ⫹ cells were found to not express any of these markers (lane
morphology of the double-labeled cells remained circular, and
5). E, GAPD expression (lanes 2–5) and proliferating marker PCNA expression (lanes 6 –9).
there were no morphological changes suggesting outer segment
development. Considering the localization of these cells in the
eration marker PCNA (data not shown). The delivered MSCs
ON and that opsin is exclusively expressed by photoreceptors, it
were quite evenly distributed in the injected area, but cell clumps
is proposed that RHOS immunoreactivity present in the ON
were occasionally observed (Fig. 4 E, arrows). The injected
demonstrates the presence of immature photoreceptors.
CD90 ⫹ MSCs were unable to migrate outside the perimeter of
One of the most important prerequisites for the development of
Kicic et al. • Marrow Stromal Cell Differentiation into Photoreceptors
7746 • J. Neurosci., August 27, 2003 • 23(21):7742–7749
Transduction of MSCs
To avoid rejection related to allogeneic transplantation, we propose autologous heterotopic transplantation using the recipients’
own MSCs. The genetic abnormality responsible for the development of retinal dystrophy or degeneration will therefore still be
present in the MSCs. To investigate the potential of combining
gene and stem cell therapies, we used a rAAV.GFP to transduce
solely expressing CD90 ⫹ MSCs. In our hands, ⬃5–7% of CD90 ⫹
MSCs were successfully transduced (Fig. 6 A, B). GFP ⫹ cells were
sorted and expanded once for subsequent in vivo experiments. At
3 weeks after injection, examination of whole-mounted retinas
transplanted with rAAV.GFP-transduced CD90 ⫹ MSCs showed
the presence of GFP ⫹ cells covering approximately one-third of
the total retina (Fig. 6C). The signal intensity decreased with
increasing distance from the injection site (Fig. 6 D) and finally
disappeared beyond the border of the retinal bleb. Histological
assessment of the injection site (Fig. 6 E, F ) confirmed this finding
with a large proportion of GFP ⫹ cells located at the site of injection (Fig. 6 E, arrows) within the ON (Fig. 6 F, arrows).
Discussion
Figure 4. Ophthalmic examination of CD90 ⫹ MSC transplanted retinas at different time
intervals. A–H, Fundus photographs ( A–F) and fluorescein angiograms (G, H ) after subretinal
injection of CD90 ⫹ MSCs. Subretinal injection induced temporary retinal detachment ( A, B,
circled by dotted lines) with the size of one-third of the whole retina. C–F, Distribution of
CD90 ⫹ MSCs at 2 weeks (C, D) and 5 weeks ( E, F ) after injection. Even distribution with some
clumped cells is present (arrows). Note that no teratoma formation was observed up to 5 weeks
after injection. G, H, Fluorescein angiograms of CD90 ⫹ MSCs ( G) and vehicle (control) ( H )injected eyes at 5 weeks after injection. Note that no sign of ocular neovascularization is present.
Asterisks indicate sites of subretinal injection. Solid circles, Artifactual reflex from the modified
clinical fundus camera.
a functional photoreceptor layer is the ability of the transplanted cells
to establish neuronal connections. Synaptophysin, a synaptic vesicle
protein involved in neuronal transmission (Weidmann and Franke,
1985), is typically observed in the outer plexiform and ganglion layers of normal retina (Fig. 5P). Immunohistochemical studies revealed that the BrdU-positive cells in CD90 ⫹ MSC-injected eyes
colocalized within areas that were immunoreactive for synaptophysin (Fig. 5Q). Because a majority of BrdU-labeled cells was observed
primarily at the proximal end of the outer nuclear layer by 5 weeks
after transplantation (Fig. 5R), closer inspection of these sites via
electron microscopy demonstrated the presence of synaptic vesicles
surrounding transplanted cells (Fig. 5S).
Embryonic stem cell therapies may hold enormous potential for
the treatment of a wide range of degenerative diseases. However,
this approach remains highly controversial, and in many countries, legislation has been introduced to limit their derivation and
use (Vogel, 2000). In addition, the problem of rejection still needs
to be resolved. This has been highlighted by a recent study by
Drukker and colleagues (2002), which has shown that human
embryonic stem cell major histocompatibility complex class I
expression moderately increases with in vitro and in vivo differentiation, suggesting that these cells may be rejected during transplantation. In this respect, the identification of an easily available
stem cell source would offer the advantage of autologous stem cell
therapies. Results of recent investigations (Azizi et al., 1998; Brazelton et al., 2000; Mezey et al., 2000) have demonstrated the
plasticity of MSCs into neural lineages. This prompted us to attempt their differentiation into photoreceptors for both in vitro,
whereby MSCs were exposed to particular agents described previously to induce stem cell differentiation into photoreceptors,
and in vivo, whereby MSCs were transplanted into the subretinal
space of normal rats, and cell fate was assessed over 3 months.
A number of previous investigations has demonstrated MSC
plasticity by their differentiation into chondrogenic, adipogenic,
and osteogenic cell lineages (Mackay et al., 1998; Kopen et al.,
1999; Pittenger et al., 1999; Muraglia et al., 2000). This multilinear potential has been highlighted also with several recent studies
showing successful differentiation of MSCs into muscle, glia, epithelia, and liver cells (Eglitis and Mezey, 1997; Ferrari et al., 1998;
Kopen et al., 1999; Petersen et al., 1999; Orlic et al., 2001; Jiang et
al., 2002).
Preliminary characterization confirmed previous findings
that heterogenous rat MSC populations primarily express CD90
with little or no lymphohematopoietic cell marker expression
(Table 1). Analysis also revealed that these cells expressed very
low or negligible levels of CD11b and CD45, which are typically
expressed by the default cell types for MSCs. Future immunoblotting experiments on those cells that did not express any of the
marker tested in this study could then be used to obtain a complete phenotypic profile of MSCs.
A number of previous investigations have demonstrated that
taurine, activin A, and EGF are present in the retina and are
possibly involved in inducing photoreceptor differentiation both
during embryonic development (Lillien and Cepko, 1992; Davis
Kicic et al. • Marrow Stromal Cell Differentiation into Photoreceptors
J. Neurosci., August 27, 2003 • 23(21):7742–7749 • 7747
ation rather than differentiation in vitro
(Ahmad et al., 1998, 1999). However, to
our knowledge, this is the first investigation to show that these agents are capable
of inducing adult-derived CD90 ⫹ MSCs
to express photoreceptor-specific markers
in vitro. Having demonstrated their differentiation via specific marker expression by
Western blot analysis and RT-PCR, additional studies using a more quantitative
approach would be required to see whether
the specific photoreceptor marker expression patterns observed in this study correlate
to the expression patterns of normal functioning photoreceptors.
Previous studies have shown that bone
marrow-derived stem cells (Tomita et al.,
2002) and neural progenitor cells (Nishida
et al., 2000; Young et al., 2000; Pressmar et
al., 2001) can integrate and differentiate
more readily in diseased retinas than in
healthy retinas. One of the major concerns
of stem cell or pluripotent cell injection is
adverse effects related to uncontrolled cell
proliferation. Considering the potential
applications, it is an important observation that CD90 ⫹ MSCs can be successfully
injected into animals without any serious
pathological consequences. Our findings
suggest that once in the retinal microenvironment, CD90 ⫹ MSCs either undergo
differentiation into potentially functioning photoreceptors or remain undifferentiated with little disruption to the host
retinal morphology. Biological complications like that related to neoplasia seem
unlikely, because color fundus photography 12 weeks after transplantation (data
Figure 5. A–S, CD90 ⫹ MSC fate 1–12 weeks after transplantation. A, CD90 ⫹ MSCs were prelabeled with BrdU before not shown) showed a similar morphology
injection (400⫻). B–E, H, Hematoxylin and eosin ( H, E) staining of retinas 7 d after transplantation ( B–D) revealed a large to that observed at 5 weeks (Fig. 4), and
number of CD90 ⫹ MSCs in the subretinal space at the site of injection (B, arrow) (100⫻), with only slight disruption to the host
transplanted cells did not show any immuretina (C, arrow) (200⫻) occurring at the site of injection. Retina distant to the injection site ( D) (200⫻) revealed normal
structuralintegrity.E,Retinas5weeksaftertransplantationhistologicallyappearednormalwithverylittledisruptiontothehost(200⫻). noreactivity to PCNA, a proliferation
PocketsofCD90 ⫹ MSCswereobservedstillinthesubretinalspace,althoughthesewereobservedprimarilyatthesiteofinjection(arrow) marker (data not shown). However, in this
(200⫻).G, I,DetectionofBrdU-labeledCD90 ⫹ MSCsintheONatthesiteofinjectionat7d(G)(300⫻)and5weeks(I)aftertransplan- study, the signal intensity of BrdU-labeled
tation (300⫻). F, H, Retina distant to the injection site appeared normal and did not contain any transplanted cells at 7 d (F) (300⫻) and transplanted cells did in fact diminish over
5 weeks (H) (300⫻) after transplantation. Even 5 weeks after transplantation, CD90 ⫹ cells were still primarily found within the ON (J) time. This observation may suggest that in
(200⫻), adjacent to the injection site of host retinas. K, Expression of the photoreceptor-specific marker opsin revealed typical outer the case of syngenic transplantations, the
segment staining in nontransplanted retinas (300⫻). L, A similar pattern of opsin expression in host retina distant to the injection site long-term survival of transplanted cells in
(100⫻). M, At the injection site, the outer segment staining becomes less apparent, because the structural hierarchy of the host retina is the retina remains uncertain, which could
not fully restored after transplantation but is primarily expressed in the cytoplasm of cells located within the injection site (100⫻). N, be a cause for concern (Zhang and Bok,
Higher magnification of the injection site revealed the cytoplasmic expression of opsin in these cells (400⫻). O, Transplanted cells were
1998). Additional studies are required to
identified by the coexpression of BrdU (green) and opsin (red) within the injection site (1000⫻; inset, 2000⫻). Bottom arrows highlight
two injected MSCs that were immunoreactive for BrdU and the RHOS antibodies. The top arrow indicates an endogenous photoreceptor investigate this further.
Previously, Chacko et al. (2000) showed
whose opsin expression had been translocated to its cytoplasm as a result of the injection procedure. P, Synaptophysin expression in
nontransplanted retina typically stains positive (indicated as red) in the outer plexiform and ganglion cell layers (100⫻). Q, Higher that embryonic retinal progenitor cells
magnificationrevealsthecolocalizationofBrdUandsynaptophysininCD90 ⫹ MSC-transplantedretina(1000⫻).R,Immunohistochem- could be transplanted into the subretinal
ical analysis of transplanted retina revealed that the majority of BrdU-labeled cells was located at the proximal end of the ON, adjacent to space of 2-week-old rats, and that at 26 d
theOPL(1000⫻).S,Electronmicroscopicanalysisofthesecellsrevealedsynapticvesicleformation(arrows)(31,200⫻).G,Ganglioniccell after transplantation, injected progenitor
layer; IN, inner nuclear layer; OPL, outer plexiform layer; OS outer segments.
cells had incorporated into the host retina
and expressed opsin and recoverin without disrupting the morphology and lamiet al., 2000) and postnatally (Anchan et al., 1991; Altshuler et al.,
nar organization of the host retina. Work conducted by Kopen et
1993). It has been observed that embryonically derived retinal
al. (1999) observed that CD90 ⫹ MSCs could also be successfully
precursors require the presence of EGF to sustain their undifferentiated state and that its primary activity is to promote proliferengrafted into the neonatal brain, which results in their differen-
Kicic et al. • Marrow Stromal Cell Differentiation into Photoreceptors
7748 • J. Neurosci., August 27, 2003 • 23(21):7742–7749
different retinal diseases. In humans, the macula is responsible
for 80% of the vision that enables us to function with ease in
society. The treatment of human blinding conditions using this
therapeutic approach would then require concentrated functional cell rescue of this small area. Besides the primate, no other
animal model (including the rodent) possesses a macular, and
thus studying the effects of a relevant functional rescue would be
extremely difficult, because most existing biological functional
tests for rodents (ERG, behavioral tests) require rescue in a relatively large retinal surface. However, the experiments described
in this paper are highly relevant because they establish the following principles: (1) bone marrow pluripotent cells can be differentiated into photoreceptors and genetically modified to correct the
underlying disease-causing mutation, and (2) transplanted cells
survive without rejection in the retinal space, do not proliferate,
express photoreceptor-specific markers, and are capable of attracting synaptic vesicles. Having established these principles, the
next step would be to conduct primate experiments focusing on
the macula that could then potentially lead to human trials.
References
Figure 6. A, B, Phase contrast ( A) and fluorescent ( B) micrographic images of CD90 ⫹ MSCs
transduced with rAAV.GFP before transplantation. Fluorescent micrographic images of wholemounted retina at 3 weeks after transplantation of rAAV. Arrows highlight a particular CD90⫹
MSC that had been successfully transduced with rAAV-GFP. C, D, GFP-transduced CD90 ⫹ MSCs
adjacent ( C) and distant ( D) to the injection site (400⫻). E, F, Hematoxylin and eosin staining
( E; arrow indicates site of injection) and fluorescent microscopy ( F; top arrow indicates site of
injection and bottom arrow indicates transplated cells located at the site of injection) of transplanted rAAV.GFP transduced CD90 ⫹ MSCs (200⫻). G, Ganglionic cell layer; IN, inner nuclear
layer; OS, outer segments; R, retinal pigment epithelium.
tiation into astrocytes and neurons without disrupting the surrounding tissue. These observations demonstrated that photoreceptors could potentially be produced from embryonic retinal
progenitor cells, and that engrafting into neuronal tissue does not
induce neoplasia. However, the clinical benefits of these observations for the treatment of retinal degenerative disorders are
somewhat limited because of ethical issues, limited accessibility,
number of retinal progenitor cells, and the potential of tissue
rejection by the host.
Our observation that adult CD90 ⫹ MSCs injected into the
subretinal space of adult rats can be differentiated into photoreceptors with similar efficiency as embryonic retinal progenitor
cells offers a potential treatment strategy for retinal degenerations
and dystrophies. Our in vivo results also correlate well to the
preliminary findings of Tomita and colleagues (2002), who
showed similar levels of integration and differentiation of bone
marrow-derived stem cells when injected into the intravitreous
space of rats whose retinas were mechanically injured. Importantly, we found that adult autologous heterotopic transplantation of CD90 ⫹ MSCs recovered from the recipient bone marrow
and corrected for the genetic abnormality are capable of becoming functional photoreceptors, thus restoring dysfunctional retinas in humans. However, this study represents only the first step
in this path, and several questions regarding the ability of the
differentiated MSCs to develop fully functional rod and cone
receptors remain to be answered.
The ultimate aim of this research is to rescue function in the
macula, or more precisely the cone rich foveola, of sufferers of
Ahmad I, Dooley CM, Afiat S (1998) Involvment of Mash1 in EGFmediated regulation of differentiation in the vertebrate retina. Dev Biol
194:86 –98.
Ahmad I, Dooley CM, Thoreson WB, Rogers JA, Afiat S (1999) In vitro
analysis of a mammalian retinal progenitor that gives rise to neurons and
glia. Brain Res 831:1–10.
Altshuler D, Lo Turco JJ, Rush J, Cepko C (1993) Taurine promotes the
differentiation of a vertebrate retinal cell type in vitro. Development
119:1317–1328.
Anchan RM, Reh TA, Angello J, Balliet A, Walker M (1991) EGF and TGFalpha stimulate retinal neuroepithelial cell proliferation in vitro. Neuron
6:923–936.
Azizi SA, Stokes D, Augelli BJ, DiGirolamo C, Prockop DJ (1998) Engraftment and migration of human bone marrow stromal cells implanted in
the brains of albino rats–similarities to astrocyte grafts. Proc Natl Acad Sci
USA 95:3908 –3913.
Besharse JC, Pfenninger KH (1980) Membrane assembly in retinal photoreceptors I. Freeze-fracture analysis of cytoplasmic vesicles in relationship
to disc assembly. J Cell Biol 87:451– 463.
Brazelton TR, Rossi FM, Keshet GI, Blau HM (2000) From marrow to brain:
expression of neuronal phenotypes in adult mice. Science 290:1775–1779.
Chacko DM, Rogers JA, Turner JE, Ahmad I (2000) Survival and differentiation of cultured retinal progenitors transplanted in the subretinal space
of the rat. Biochem Biophys Res Commun 268:842– 846.
Davis AA, Matzuk MM, Reh TA (2000) Activin A promotes progenitor differentiation into photoreceptors in rodent retina. Mol Cell Neurosci
15:11–21.
Drukker M, Katz G, Urbach A, Schuldiner M, Markel G, Itskovitz-Eldor J,
Reubinoff B, Mandelboim O, Benvenisty N (2002) Characterization of
the expression of MHC protein in human embryonic stem cells. Proc Natl
Acad Sci USA 99:9864 –9869.
Eglitis MA, Mezey E (1997) Hematopoietic cells differentiate into both microglia and macroglia in the brains of adult mice. Proc Natl Acad Sci USA
94:4080 – 4085.
Ferrari G, Cusella-De Angelis G, Coletta M, Paolucci E, Stornaiuolo A, Cossu
G, Mavilio F (1998) Muscle regeneration by bone marrow-derived myogenic progenitors. Science 279:1528 –1530.
Gavrieli Y, Sherman Y, Ben-Sasson SA (1992) Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 119:493–501.
Huang JC, Ishida M, Hersh P, Sugino IK, Zarbin MA (1998) Preparation
and transplantation of photoreceptor sheets. Curr Eye Res 17:573–585.
Jiang Y, Jahagirdar BN, Reinhardt RL, Schwartz RE, Keene CD, OrtizGonzalez XR, Reyes M, Lenvik T, Lund T, Blackstad M, Du J, Aldrich S,
Lisberg A, Low WC, Largaespada DA, Verfaillie CM (2002) Pluripotency of mesenchymal stem cells derived from adult marrow. Nature
418:41– 49.
Kopen GC, Prockop DJ, Phinney DG (1999) Marrow stromal cells migrate
Kicic et al. • Marrow Stromal Cell Differentiation into Photoreceptors
throughout forebrain and cerebellum, and they differentiate into astrocytes after injection into neonatal mouse brains. Proc Natl Acad Sci USA
96:10711–10716.
Lai CM, Brankov M, Zaknich T, Lai YK, Shen WY, Constable IJ, Rakoczy PE
(2001) Inhibition of angiogenesis by adenovirus-mediated sFlt-1 expression in a rat model of corneal neovascularization. Hum Gene Ther
12:1299 –1310.
Lillien L, Cepko C (1992) Control of proliferation in the retina: temporal
changes in responsiveness to FGF and TGF alpha. Development
115:253–266.
Lolley RN, Rong H, Craft CM (1994) Linkage of photoreceptor degeneration by apoptosis with inherited defect in phototransduction. Invest Ophthalmol Vis Sci 35:358 –362.
Mackay AM, Beck SC, Murphy JM, Barry FP, Chichester CO, Pittenger MF
(1998) Chondrogenic differentiation of cultured human mesenchymal
stem cells from marrow. Tissue Eng 4:415– 428.
Mezey E, Chandross KJ, Harta G, Maki RA, McKercher SR (2000) Turning
blood into brain: cells bearing neuronal antigens generated in vivo from
bone marrow. Science 290:1779 –1782.
Muraglia A, Cancedda R, Quarto R (2000) Clonal mesenchymal progenitors from human bone marrow differentiate in vitro according to a hierarchical model. J Cell Sci 113:1161–1166.
Nishida A, Takahashi M, Tanihara H, Nakano I, Takahashi JB, Mizoguchi A,
Ide C, Honda Y (2000) Incorporation and differentiation of
hippocampus-derived neural stem cells transplanted in injured rat retina.
Invest Ophthalmol Vis Sci 41:4268 – 4274.
Orlic D, Kajstura J, Chimenti S, Jakoniuk I, Anderson SM, Li B, Pickel J,
McKay R, Nadal-Ginard B, Bodine DM, Leri A, Anversa P (2001) Bone
marrow cells regenerate infarcted myocardium. Nature 410:701–705.
Petersen BE, Bowen WC, Patrene KD, Mars WM, Sullivan AK, Murase N,
Boggs SS, Greenberger JS, Goff JP (1999) Bone marrow as a potential
source of hepatic oval cells. Science 284:1168 –1170.
Phinney DG, Kopen G, Isaacson RL, Prockop DJ (1999) Plastic adherent
stromal cells from the bone marrow of commonly used strains of inbred
mice: variations in yield, growth, and differentiation. J Cell Biochem
72:570 –585.
J. Neurosci., August 27, 2003 • 23(21):7742–7749 • 7749
Pittenger MF, Mackay AM, Beek SC, Jaiswal RK, Douglas R, Mosca JD, Moorman MA, Simonetti DW, Craig S, Marshak DR (1999) Multilineage potential of adult human mesenchymal stem cells. Science 284:143–147.
Pressmar S, Ader M, Richard G, Schachner M, Bartsch U (2001) The fate of
heterotopically grafted neural precursor cells in the normal and dystrophic adult mouse retina. Invest Ophthalmol Vis Sci 42:3311–3319.
Shen WY, Rakoczy PE (2001) Uptake dynamics and retinal tolerance of
phosphorothioate oligonucleotide and its direct delivery into the site of
choroidal neovascularization through subretinal administration in the
rat. Antisense Nucleic Acid Drug Dev 11:257–264.
Tomita M, Adachi Y, Yamada H, Takahashi K, Kiuchi K, Oyaizu H, Ikebukuro K, Kaneda H, Matsumura M, Ikehara S (2002) Bone marrowderived stem cells can differentiate into retinal cells in injured rat retina.
Stem Cells 20:279 –283.
Tropepe V, Coles BL, Chiasson BJ, Horsford DJ, Elia AJ, McInnes RR, van der
Kooy D (2000) Retinal stem cells in the adult mammalian eye. Science
287:2032–2036.
Vogel G (2000) Stem cells: new excitement, persistent questions. Science
290:1672–1674.
Wiedenmann B, Franke WW (1985) Identification and localization of synaptophysin, an integral membrane glycoprotein of Mr 38,000 characteristic of presynaptic vesicles. Cell 41:1017–1028.
Woch G, Aramant RB, Seiler MJ, Sagdullaev BT, McCall MA (2001) Retinal
transplants restore visually evoked responses in rats with photoreceptor
degeneration. Invest Ophthamol Vis Sci 42:1669 –1676.
Woodbury D, Schwarz EJ, Prockop DJ, Black IB (2000) Adult rat and human bone marrow stromal cells differentiate into neurons. J Neurosci Res
61:364 –370.
Young MJ, Ray J, Whiteley SJ, Klassen H, Gage FH (2000) Neuronal differentiation and morphological integration of hippocampal progenitor cells
transplanted to the retina of immature and mature dystrophic rats. Mol
Cell Neurosci 16:197–205.
Zhang X, Bok D (1998) Transplantation of retinal pigment epithelial cells
and immune response in the subretinal space. Invest Ophthalmol Vis Sci
39:1021–1027.